An All-data-segment Radio Frequency Fingerprint Extraction Method Based on Cross-power Spectrum
Dawei Fang, Aiqun Hu, Junxian Shi · 2022 2nd International Conference on Consumer Electronics and Computer Engineering (ICCECE) · 2022
Radio frequency fingerprint (RFF) is the inherent signal characteristic of wireless devices and has been employed for identification. Its stability and robustness against noise has brought the necessity of further research for practical use. This paper proposes a novel RFF extraction method based on cross-power spectrum. This method could extract RFF from all signal segments, i.e., the device could send arbitrary content in data segments for RFF extraction rather than repetitive training symbols. Experimental verifications used 50 CC2540 ZigBee modules for classification. The result shows that the proposed scheme can achieve a 97% accuracy rate at 20dB signal-to-noise-rate (SNR), and performs more efficiently than other anti-noise methods in low SNR scenarios.